5102
J . A m . Chem. SOC.1982, 104, 5102-5105
due to this complex. In the present study this was accomplished by thorough electrochemical or chemical reduction to obtain pure solutions of the [V(3,5-di-rert-b~tylcatecholate)~]~ion. It should be noted that the extinction coefficient reported by Sawyer et al. is less than half of the correct value for [V(3,5-di-tert-butylcatecholate),]-; this indicates the extent of oxidation of the V(1V) starting material and implies that over half of the vanadium remained in the tetravalent state to give rise to the EPR spectrum ( g = 1.98, A = 107.8 G) observed by Sawyer et al. It is likely that this EPR spectrum is due to unreacted VO(acac)z, which has g = 1.969 and A = 105 G in the same solvent. Thus the data reported by Sawyer et represent the optical spectrum of the V(V) complex with the E P R spectrum of a V(1V) complex. In does not the absence of base, [V(3,5-di-tert-butyl~atecholate)~]~form in MeOH (similar results are obtained for catechol); preI- does form because sumably [V(3,5-di-tert-b~tylcatecholate)~] its higher metal ion charge leads to a higher stability constant, which permits it to displace the catechol protons without need for exogenous base. More particularly, we have been able to identify the putative “oxygen adduct” as a vanadium(V) catechol complex. This complex is formed by reaction with N O or 02,although not stoichiometrically in the former case; we have found that purple solutions generated by introduction of N O exhibit an EPR spectrum with two sets of hyperfine couplings ( A = 107 and 97 X 10“ cm-I), which are characteristic of vanadyl acetylacetonate and mono(catecho1ate). Thus the ”complicated EPR spectrum that is consistent with the patterns that have been observed for square-pyramidal geometry”,21which Sawyer et al. attributed to superhyperfine splitting from the nitrogen of an axially symmetric N O ligand, is in fact due to a mixture of simple vanadyl complexes.
We have not in any of our experiments observed generation of the purple V(V) complex by exposure to pure C O as reported by Sawyer et al. and suggest that the improbable observation of superimposable optical spectra for both CO and O2adducts is more reasonably explained by exposure to oxygen in both cases. Some of the more puzzling aspects of the report by Sawyer et al. remain unexplained-in particular the “reversibility” of “adduct formation”, which we have not been able to reproduce. However, it is clear that the observations of Sawyer et al. are not due to reversible oxygen binding by vanadium catecholate complexes.
Acknowledgment. We wish to acknowledge the support of the N I H through Grant AI 11744. S.R.C. wishes to acknowledge the donors of the Petroleum Research Fund, administered by the American Chemical Society, for partial support of this research and to thank Professor E. J. Corey and K. Blwh of this department for use of their spectrophotometers. We also thank Dr. Frederick J. Hollander for his able experimental assistance and instruction in the structure analysis. Funds for the U.C.B. Chexray facility were provided by the NSF. Registry No. [Et3NH]2[V(cat)3]CH3CN, 82613-78-3; [Et3NH][V(DTBC),], 82613-80-7; [Et,NHJ,[VO(DTBC),], 82613-82-9;K2[VO(~at)~l.EtOH.H~0, 82659-76-5; K3[V(cat)3].l.5H20,82613-84-1;VO(acac),, 3153-26-2; NH4[V03],7803-55-6; [V(DTBC),I2-,82613-85-2; [V(DTBC),]-,82613-86-3; [V(cat)J, 82613-87-4; VO(cat), 82598-73-0; [VO(Tironate)12-,82613-88-5; VO(DTBC), 82598-74-1; [VO(Tironate),]&, 82621-18-9; [V(Tironate),]*-,82613-89-6. Supplementary Material Available: Listing of observed and calculated structure factors (60 pages). Ordering information is given on any current masthead page.
Angle-Resolved Ultraviolet Photoelectron Spectroscopic Studies of CO Binding to Three Chemically Different Surfaces of ZnO. Confirmation of Step-Binding Sites on
(oooi) K. L. D’Amico, M. Trenary, N. D. Shinn, Edward I. Solomon,*’ and F. R. McFeely* Contribution f r o m the Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139. Received December 7, 1981
Abstract: Angle-resolved photoelectron spectroscopy (ARPES) has been performed on the Zn0(000i)-CO system. These measurements are compared with previous ARPES results on (1010) and (0001) surfaces and strongly support earlier suggestions that the CO which is observed to chemisorb to this surface binds to (1010) step sites which contain coordinately unsaturated zinc ions but not to the (0001) terrace sites which contain only coordinatively unsaturated oxide ions.
The adsorption of C O on ZnO is an important problem in relation to methanol synthesis.z The CO-ZnO bond is unusual because of the observation of a -70-cm-’ increase in the C O stretching frequency upon chemisorption of CO on ZnO powders., This is in contrast to the decrease in the stretching frequency relative to the gas-phase value of 2143 cm-I, which is normally (1) Present address: Department of Chemistry, Stanford University, Stanford, CA 94305. (2) A. L. Waddams, “Chemicals from Petroleum”,3rd ed., Wiley, New York, 1973. (3) (a) F. Boccuzzi, G . E. Garrone, A. Zecchina, A. Bossi, and M. Camia, J . Cutul., 51, 160 (1978); (b) N. S. Hush and M. L. Williams, J . Mol. Spectrosc., 50, 349 (1974).
0002-7863/82/1504-5102$01.25/0
observed upon chemisorption to metals and in organometallic complexes. In an effort to gain a detailed microscopic understanding of the nature of the CO/ZnO interaction, we have recently reported a variety of ultraviolet photoelectron spectroscopic (UV PES) studies on four chemically different surfaces of ZnO:@ (OOOl), which contains only coordinatively unsaturated zinc sites; (1010) and (1 120), which contain both zinc and oxide ions with (4) R. R. Gay, M. H. Nodine, E. I. Solomon, V. E. Henrich, and H. J. Zeiger, J . Am. Chem. SOC.,102, 6752 (1980). (5) M. J. Sayers, M. R. McClellan, R. R. Gay, E. I. Solomon, and F. R. McFeely, Chem. Phys. Left., 75, 575 (1980). (6) M. R. McClellan, M. Trenary, N. D. Shinn, M. J. Sayers, K. L. DAmico, E. I. Solomon, and F. R. McFeely, J. Chem. Phys., 74,4726 (1981).
0 1982 American Chemical Society
J . Am. Chem. Sot., Vol. 104, No. 19, 1982 5103
Confirmation of Step-Binding Sites on (0007)
[oooTI
0
0
0
0
02P
Zn3d I
C040
I
0
[roio]
I
I
0
I
0
I
I
0
0
/zl"\ Zn4bZn/z:n\ Zn/in\ Zn/zln\ Zn
Zn
I
I
I
I
0
0
0
I
(zln\Zn
Zn
I
I
Figure 1. Structure of the ZnO (0001) surface, illustrating the doublelayer step site known to exist on this surface. Such a step exposes a (1010) type site onto which a C O could bind to a coordinatively unsaturated Zn ion. The Zn-C-0 surface complex is oriented parallel to the [ lOlO] direction which is approximately 20° out of the plane of the paper. Due to the symmetry of the (0001) surface, this site exists in six equivalent domains.
coordinatively unsaturated positions; (OOOi), which nominally contains only oxide ions with open coordination positions directed normal to the surface. These studies have clearly demonstrated that CO binds carbon-end down4to the coordinatively unsaturated zinc site on the (OOO1) and (1070)surfaces, forming approximately linear Zn-C-0 complexes. Here C O acts as a Sa donor by transferring electron density from this weakly antibonding orbital, with little A back-bonding to the CO 2 ~ orbital, * resulting in a net donation of charge to the surface zinc ion. Hence the C O bond order is increased, and the CO molecule is polarized with some positive charge a t the carbon. A puzzling aspect of these studies, however, was the observation4 of a small but nonzero amount of C O 40 photoelectron intensity indicative of C O bonding to the (OOOT) surface, which ideally contains only oxide ions. The temperature and pressure dependence of this 40 intensity further indicated that the CO binds with the same heat of adsorption at zero coverage to (000T)as to the other three surfaces. This led to the proposal that C O binds to zinc-containing step-defect sites on this surface. LEED studies' on (OOOT) indicate a significant density of (lOT0) steps, which expose zinc sites with coordinatively unsaturated positions -70' off the surface normal (Figure 1). In this paper we present angle-resolved UV PES data that demonstrate that C O binds with an angle a of -70' with respect to the surface normal, strongly supporting the proposal that C O binds t o these step sites.
Experimental Section The crystal used in these experiments was cut from a needle of vapor-phase-grown material and was oriented by Laue diffraction to within A I o . The surface was then polished with 1 pm Al2O3and etched with dilute HC1 to produce a surface free of etch pits as detectable by optical microscopy. The surface was cleaned in situ by repeated cycles of mild (
